Actin Accelerates Plasmin Generation by Tissue Plasminogen Activator*

Actin has been found to bind to plasmin’s kringle regions, thereby inhibiting its enzymatic activity in a noncompetitive manner. We, therefore, examined its effect upon the conversion of plasminogen to plasmin by tissue plasminogen activator. Actin stimulated plas- min generation from both Glu- and Lys-plasminogen, lowering the K , for activation of Glu-plasminogen into the low micromolar range. Accelerated plasmin generation did not occur in the presence of t-amino caproic acid or if actin was exposed to acetic anhydride, an agent known to acetylate lysine residues. Actin binds to tissue plasminogen activator (t-Pa) ( K d = 0.55 pM), at least partially via lysine-binding sites. Actin’s stim- ulation of plasmin generation from Glu-plasminogen was inhibited by the addition of aprotinin and was restored by the substitution of plasmin-treated actin, indicating the operation of a plasmin-dependent positive feedback mechanism. Native actin binds to Lys- plasminogen, and promotes its conversion to plasmin even in the presence of aprotinin, indicating that plas- min’s cleavage of either actin or plasminogen leads to further plasmin generation. Plasmin-treated actin binds Glu-plasminogen and t-PA simultaneously, thereby raising the local concentration of t-PA and plasminogen. Together, but not separately, actin and t-PA prolong the thrombin A fresh aliquot thawed immediately before each experiment. Kinetic Analysis of Plasmin Generation and t-PA Actiuity-Kinetic measurements were performed at 21 "C in wells of polystyrene micro- titer plates (1.0 X 0.6 cm) using a Bio-Tek Instruments enzyme-linked immunosorbent assay plate reader to monitor liberation of p- nitroaniline from S-2288 (H-D-Ile-Pro-Arg-pNA) or S-2251 (H-D-Val-Leu-Lys-pNA) at 405 nm. Conversion of plasminogen to plasmin was monitored in a two-stage assay similar to that described by Hoylaerts et al. (24) except that 10 mM e-ACA was included in the second stage to prevent inhibition of the plasmin generated by actin (21). In the first stage, Glu-plasminogen, t-PA, and varying amounts of e-ACA, actin, CNBr cleavage products of fibrinogen, or soluble fibrin were incubated in 40 pl of a solution containing 20 mM Tris, 150 mM NaC1, pH 7.4 (TBS), at 37 "C. In the second stage, 5-pl aliquots of the incubation mixture were added to wells of a microtiter plate that contained 95 ml of a TBS solution containing S-2251 (0.2 mM) and e-ACA (10 mM), thereby diluting the primary reactants by a factor of 20. The initial rate of change of the absorbance of the solution at 405 nm/min of the stage I incubation time was used to determine the rate of plasmin generation by using a standard curve constructed by determining the rate of S-2251 hydrolysis by known quantities of plasmin. actin,

Actin has been found to bind to plasmin's kringle regions, thereby inhibiting its enzymatic activity in a noncompetitive manner. We, therefore, examined its effect upon the conversion of plasminogen to plasmin by tissue plasminogen activator. Actin stimulated plasmin generation from both Glu-and Lys-plasminogen, lowering the K , for activation of Glu-plasminogen into the low micromolar range. Accelerated plasmin generation did not occur in the presence of t-amino caproic acid or if actin was exposed to acetic anhydride, an agent known to acetylate lysine residues. Actin binds to tissue plasminogen activator (t-Pa) ( K d = 0.55 p M ) , at least partially via lysine-binding sites. Actin's stimulation of plasmin generation from Glu-plasminogen was inhibited by the addition of aprotinin and was restored by the substitution of plasmin-treated actin, indicating the operation of a plasmin-dependent positive feedback mechanism. Native actin binds to Lysplasminogen, and promotes its conversion to plasmin even in the presence of aprotinin, indicating that plasmin's cleavage of either actin or plasminogen leads to further plasmin generation. Plasmin-treated actin binds Glu-plasminogen and t-PA simultaneously, thereby raising the local concentration of t-PA and plasminogen. Together, but not separately, actin and t-PA prolong the thrombin time of plasma through the generation of plasmin and fibrinogen degradation products. Actin-stimulated plasmin generation may be responsible for some of the changes found in peripheral blood following tissue injury and sepsis.
Plasminogen is a plasma glycoprotein of 791 amino acids that is converted to the active enzyme plasmin by the cleavage of the Arg""-VaPG1 peptide bond (1). The first 76 amino acids comprise the preactivation peptide, which may be cleaved by plasmin at L Y S~~, thereby converting native Glu-plasminogen to Lys-plasminogen. Five kringle domains follow, each of which is approximately 90 amino acids in length, ending at position 542. These regions mediate the binding of plasminogen to other proteins, such as fibrin, by binding to lysine residues presented in the appropriate orientation. Cleavage of the Arg5G0-Va15G' peptide bond by plasminogen activators such as tissue-plasminogen activator or urokinase results in the * This study was supported by Grant HL 42457 from the National Institutes of Health, Department of Health and Human Services. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed: Div. of Experimental Medicine, Brigham and Women's Hospital, 75 Francis St., Boston, MA 02115. generation of the active enzyme plasmin, whose enzymatically active site is comprised of residues and Ser741.

( 2 , 3 )
Actin, one of the most abundant cellular proteins, is found in all eukaryotic cells (4). It has also been found in micromolar concentrations in the circulation of healthy subjects (5-7) and in the blood of humans and animals experiencing various types of tissue injury (8-13). Extracellular actin is presumed to be released from dying cells, but the observation that certain isoforms of actin are released by viable cultured myoblasts (14) suggests that a fraction of the actin in the circulation may be purposefully secreted by healthy cells. Introduction of actin into plasma results in its binding to two high affinity actin-binding proteins, vitamin D-binding protein (also known as Gc globulin) and plasma gelsolin (15, 16). Together, these proteins constitute a defense system designed to shorten actin filaments and clear actin from the circulation, thereby protecting the host from actin's deleterious effects upon the microvasculature (17-20).
Because it seems likely that fibrin clots formed at sites of tissue injury would contain trapped actin filaments, we studied the effects of actin upon clot lysis (21). We observed that actin inhibited clot lysis, whether initiated by the addition of t-PA' and plasminogen or the direct addition of plasmin. Further analysis revealed that actin binds to the kringle regions of plasmin, thereby inhibiting its cleavage of both small peptide substrates and large proteins. Given these observations, we were interested in studying the effect of actin upon plasmin generation by t-PA, a protein that itself contains two kringle regions (22).
Protein Preparations-Rabbit skeletal muscle F-actin was used in all experiments. Its preparation and modification by acetic anhydride, as well as the isolation of Glu-plasminogen and fibrinogen, were performed as previously described (21). Plasmin-treated actin was prepared by incubating 100 p1 of F-actin (10 mg/ml) with 10 pl of plasmin (1.8 p~) for 60 min at 21 "C, followed by the addition of 10 pl of aprotinin (10,000 KIU/ml). CNBr fragments of fibrinogen were prepared by the method of Verheijen et al. (23). Fresh frozen citrated ' The abbreviations used are: t-PA, tissue plasminogen activator; c-ACA, c-amino caproic acid; SDS, sodium dodecyl sulfate. plasma, prepared by the blood transfusion service, was thawed, aliquoted, and stored at -70 "C. A fresh aliquot was thawed immediately before each experiment.
Kinetic Analysis of Plasmin Generation and t-PA Actiuity-Kinetic measurements were performed at 21 "C in wells of polystyrene microtiter plates (1.0 X 0.6 cm) using a Bio-Tek Instruments enzymelinked immunosorbent assay plate reader to monitor liberation of pnitroaniline from S-2288 (H-D-Ile-Pro-Arg-pNA) or S-2251 (H-D-Val-Leu-Lys-pNA) a t 405 nm. Conversion of plasminogen to plasmin was monitored in a two-stage assay similar to that described by Hoylaerts et al. (24) except that 10 mM e-ACA was included in the second stage to prevent inhibition of the plasmin generated by actin (21). In the first stage, Glu-plasminogen, t-PA, and varying amounts of e-ACA, actin, CNBr cleavage products of fibrinogen, or soluble fibrin were incubated in 40 pl of a solution containing 20 mM Tris, 150 mM NaC1, pH 7.4 (TBS), at 37 "C. In the second stage, 5-pl aliquots of the incubation mixture were added to wells of a microtiter plate that contained 95 ml of a TBS solution containing S-2251 (0.2 mM) and e-ACA (10 mM), thereby diluting the primary reactants by a factor of 20. The initial rate of change of the absorbance of the solution at 405 nm/min of the stage I incubation time was used to determine the rate of plasmin generation by using a standard curve constructed by determining the rate of S-2251 hydrolysis by known quantities of plasmin.
Electrophoretic Analysis of Plasmin Generation-Conversion of plasminogen to plasmin by t-PA, in the presence or absence of actin or plasmin-treated actin, was determined by incubating mixtures of these proteins at 37 "C for varying periods of time, after which they were boiled in gel sample buffer (25) and electrophoresed in 5-15% SDS-polyacrylamide gels under reducing conditions.
Binding Studies-Binding of actin to plasminogen and/or t-PA was assessed by centrifuging 12-20 ml of a TBS solution containing 10-18 pg of actin (or plasmin-treated actin) and either 10 pg of t-PA or 10 pg of plasminogen (Glu-or Lys-) at 150,000 X g for 30 min a t 21 "C in a Beckman airfuge. In some cases 10 mM e-ACA was included. The resulting supernatant fluid and pellet were separated, boiled in gel sample buffer in the presence of a reducing agent, and electrophoresed on 5-15% polyacrylamide gels.
Binding of actin to t-PA was also studied by adding actin (5 pg) to t-PA (1.5-7.5 pg) in TBS (total volume, 100 pl). The resulting mixture was either assayed for t-PA activity or centrifuged a t 150,000 X g for 45 min and the pellet and supernatant fluid separated. The t-PA activity of both the uncentrifuged mixture and the supernatant fluid was measured by adding 90 pl of either to wells of a microtiter plate that contained TBS (5 p1) and S-2288 (5 pl, final concentration, 0.5 mM) and monitoring the initial rate of change of absorbance of the solution a t 405 nm. These measurements were used as a measure of the total and free t-PA, respectively. The dissociation constant was determined by Scatchard analysis.
Thrombin Time Measurements-The thrombin time of plasma samples containing varying amounts of actin and t-PA were measured with a fibrometer (Baltimore Biological Laboratories, Baltimore, MD). Citrated plasma was diluted 50% by the addition of varying amounts of t-PA, actin, and appropriate amounts of TBS. After incubation a t 37 "C, 200-pl aliquots were added to plastic cups prewarmed to 37 "C in the fibrometer. 100 pl of a thrombin solution in T B S (10 units/ml) was then added and the time required for a clot t o form measured.
Immunologic Detection of Plasmin and Fibrinogen Degradation 'Products-Actin (final concentration, 9 mM) or T B S was added to 32 pl of citrated plasma containing varying concentrations of t-PA, in a total volume of 40 pl. After incubation at 37 "C for 45 min, gel sample buffer (120 p l ) was added and the samples placed in a boiling water bath for 2 min. 6-pl aliquots were electrophoresed on 5% (antifibrinogen antibody) or 5-15% {anti-plasminogen antibody) SDSpolyacrylamide gels and transferred +.o nitrocellulose paper in preparation for immunoblotting (26).

Actin Accelerates t-PA-mediated
Plasmin Generation-Plasmin was generated in solutions containing t-PA and Gluplasminogen only after a prolonged incubation period. The addition of rabbit skeletal muscle F-actin (actin) resulted in a dose-dependent acceleration in the rate of conversion of plasminogen to plasmin, as shown in Fig. 1 (actin did not cause plasmin generation in the absence of t-PA, data not shown). The reaction obeyed Michaelis-Menten kinetics, as demonstrated when the data were examined with a Lineweaver-Burk plot (Fig. 2). Micromolar concentrations of actin progressively lowered the K , for activation but had little effect upon kcat (Table I).
The effect of actin upon plasmin generation in the presence of known accelerators of plasminogen activation was examined. Actin was added to solutions of t-PA and Glu-plasminogen containing either CNBr-cleaved fibrinogen fragments or  soluble fibrin (Fig. 3). The rate of plasmin generation in the presence of CNBr fibrinogen fragments was not altered by the addition of actin. Although higher peak plasmin concentrations were found when both actin and CNBr fragments were added, the effect was less than additive. Actin accelerated the initial rate of plasmin generation brought about by soluble fibrin, without changing the peak plasmin concentration. Actin's stimulation of t-PA-mediated plasmin generation was abrogated by the lysine analogue 6-ACA (data not shown), suggesting that lysine residues of actin were interacting with kringle structures of t-PA and/or Glu-plasminogen. Actin also promoted plasmin generation from Lys-plasminogen. Modification of actin's lysine residues by treatment with acetic anhydride abolished its stimulatory effect for both Glu-and Lys-plasminogen (data not shown).
Interaction of Actin with t-PA and Plasminogen-The binding of actin to t-PA, Glu-plasminogen, and Lys-plasminogen was examined by subjecting mixtures of these proteins to high speed centrifugation. As shown in Fig. 4, A and B

, t-PA and
Lys-plasminogen cosedimented with actin filaments, while little Glu-plasminogen bound to actin. (Neither t-PA nor Lysplasminogen sedimented in the absence of actin, data not shown,) Addition of 10 mM e-ACA partially reversed the interaction between t-PA and actin (Fig. 4B), suggesting that actin was binding to at least one of the kringle regions of t-PA. Because low concentrations of actin did not inhibit t-PA hydrolysis of S-2288, t-PA affinity for actin could be determined by measuring t-PA binding to actin filaments using a  4 ) were centrifuged a t 150,000 X g for 30 min. The resulting supernatant fluids (lanes 1 and 2) and pellets (lanes 2 and 4 ) were separated and electrophoresed on 5-15% SDS-polyacrylamide gels under reducing conditions. Panel B, effect of r-ACA upon the binding of t-PA to actin. 20-p1 solutions containing t-PA (10 pg) and actin (10 pg) were centrifuged a t 150,000 X g for 30 min in the absence (lanes 1 and 2) or presence (lanes 3 and 4 ) of 10 mM r-ACA. The resulting supernatants (lanes 1 and 3 ) and pellets (lanes 2 and 4 ) were subjected to electrophoresis on 5-15% polyacrylamide gels, as described above. Panel C, plasmin-treated actin (18 pg) was substituted for native actin in the experiment shown in panel A. All gels were stained with Coomassie Blue. HC, plasmin heavy chain. functional assay. Scatchard analysis revealed that Kd = 0.55 p M (Fig. 5).
Interactions and Effects of Plasmin-treated Actin-Given these results, we considered whether actin's acceleration of plasmin generation might depend upon plasmin's initial cleavage of one or more of the primary reactants. This idea was confirmed by demonstrating that actin did not stimulate Gluplasminogen's conversion to plasmin in the presence of the plasmin inhibitor aprotinin (Fig. 6A). Because plasmin generation proceeded when plasmin-treated actin was substituted for native actin (in the presence of aprotinin), actin's stimulatory effect appears to result from its modification by the initial plasmin molecules formed. Actin also stimulated plasmin generation from Lys-plasminogen (Fig. 6B), an effect that was also inhibited by aprotinin. Plasmin-treated actin did not reverse aprotinin's effect, however, suggesting a difference in the interaction of one-chain and two-chain t-PA in this system. Overall, these experiments indicate that actin's stimulatory effects upon plasmin generation are due to a positive feedback mechanism, mediated by plasmin's cleavage of either actin or plasminogen. Plasmin-treated actin appears to differ from native actin largely because it binds Glu-plasminogen with higher affinity (Fig. 4C). Plasmin-treated actin did not bind t-PA to any greater extent than native actin (data not shown) nor did it directly stimulate t-PA hydrolytic activity (data not shown). Glu-plasminogen did not sediment with plasmin-treated actin that was treated with carboxypeptidase B (data not shown), indicating that Glu-plasminogen binds to a newly created COOH-terminal lysine, probably that formed by plasmin's cleavage of actin at Ly~"~-Cys"~ (27). Glu-plasminogen did not displace t-PA from plasmin-treated actin filaments when solutions containing all three molecules were centrifuged (data not shown). These results suggest that plasmin-treated actin's effect is due to its ability to simultaneously bind t-PA and plasminogen, thereby increasing their local concentrations and promoting plasmin generation.
Actin Accelerates Plasminogen Activation in Plasma-To determine whether actin might stimulate plasmin generation under more physiological conditions, actin and t-PA were added to citrated plasma and incubated at 37 "C. As shown in Fig. 7, actin prolonged the thrombin time of plasma containing various amounts of t-PA in a time-and dose-dependent manner, suggesting that significant quantities of plasmin and fibrinogen split products had been generated.
Plasma samples containing actin and t-PA were subsequently studied by immunoblotting. As shown in Fig. 8, actin promoted the generation of both plasmin (detected as plasmin-a2-anti-plasmin complexes) and fibrinogen breakdown  (lanes 7-9); or 8 pg of plasmin-treated actin and aprotinin (lanes [11][12]. The mixtures were then electrophoresed on 5-15% SDS-polyacrylamide gels under reducing conditions. The gel was stained with Coomassie Blue. Panel R, as described for panel A, except that Lys-plasminogen (10 pg) was substituted for Glu-plasminogen. HC, plasmin heavy chain.
products, despite the presence of plasma actin-binding proteins and plasma protease inhibitors.

DISCUSSION
Actin has not previously been reported to interact with t-PA or plasminogen, although these interactions were anticipated by the determination that actin binds to plasmin's kringle regions (21). Because the binding of actin to plasmin results in a loss of plasmin's enzymatic activity, a property of the distal part of the plasmin molecule, it seemed possible that actin might also inhibit t-PA cleavage of plasminogen's Argoo-Va1561 peptide bond. The experiments reported here indicate that, in fact, actin accelerates plasmin generation by Actin is not alone in stimulating plasmin generation by t-PA, for fibrin (28), fibrinogen (24), CNBr cleavage fragments of fibrinogen (23), polylysine (29), and a number of denatured proteins (30) also do so. This phenomenon appears to require one or more appropriately oriented lysine residues (31, 32), which likely exert their effects by interacting with the kringle regions of plasminogen and/or t-PA.
The fibrin surface is thought to accelerate plasmin generation by bringing together t-PA and plasminogen (24, 33). Degradation of fibrin by plasmin significantly speeds up the process by generating additional fibrin-binding sites for plasminogen and t-PA (34-39). These sites are created by plasmin's cleavage of fibrin at internal lysine residues, thereby t-PA. generating COOH-terminal residues. This change allows these residues, which normally bind only to the low affinity aminohexyl-binding site of kringle 5, to bind to the "lysinebinding sites" of kringles 1-4 that mediate high affinity plasminogen-fibrin binding (40-42).
The experiments reported here show that actin accelerates plasmin generation in a similar manner. Its effects are dependent upon the appropriate three-dimensional presentation of its lysine residues which interact with kringle regions of plasminogen and t-PA. The interaction of the three depends, however, upon the action of the first plasmin molecules generated in the solution, for native actin binds poorly to Gluplasminogen. Plasmin-treated actin, on the other hand, is able to simultaneously bind Glu-plasminogen and t-PA, thereby promoting plasmin generation, even in the presence of aprotinin. Actin's acceleratory effect appears to be due to its ability to bind both t-PA and plasminogen, thereby increasing their local concentration. It thus behaves in a manner similar to that described for immobilized thrombospondin and histidine-rich glycoprotein, each of which may promote plasmin generation by t-PA, particularly if first treated with plasmin (43,44).
Actin's stimulatory effect on t-PA-mediated plasmin generation may account in part for the findings of Bicsak and Hsueh (45)) who reported that t-PA present in rat oocyte extracts was fully active in the absence of an added t-PA simulator (such as fibrin or polylysine). They attributed this finding to a dissociable factor, not otherwise characterized, which was also present in media conditioned by rat insulinoma cells. The results reported here suggest that actin may be the endogenous stimulator detected by these investigators. The relevance of this interaction remains uncertain, however, until it can be shown that actin and t-PA are present simultaneously in a relevant compartment of the oocyte.
The potential clinical significance of these findings lies in actin's ubiquity (4). It is found in virtually all eukaryotic cells and is the predominant protein constituent of vertebrate platelets, white blood cells, and most other nucleated cells. Inflammation may cause endothelial cell death and/or Weibel-Palade body secretion, resulting in locally elevated t-PA concentrations in regions where actin is released from injured cells. The simultaneous introduction of both proteins into the extracellular space may overwhelm the defenses provided by plasma protease inhibitors and plasma actin-binding proteins, leading to plasmin generation, which in turn may act upon platelets, endothelial cells, and plasma proteins. Fibrinogen split products, for example, are known to have deleterious effects upon clot formation and platelet function. The release of actin from dying cells may thus amplify the effects of the inflammatory mediators that caused cell death and its release. Extracellular actin may contribute, in some circumstances, to the clinical and laboratory findings of disseminated intravascular coagulation, a syndrome that frequently accompanies sepsis and tissue injury.